JP2006317671A - Signal processing circuit, signal processing method, imaging device, method for speech signal processing of imaging device, recording device and recording method, and reproduction device and reproduction method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce a cyclic noise of a speech signal in which a cyclic noise is included. <P>SOLUTION: A correlation calculation is implemented for one cycle of a speech input signal (S+N) inputted in an input terminal 10 and a speech input signal of one cycle before being supplied from a buffer memory 11 at a correlation calculator 12, and a correlation coefficient C is calculated. A gain coefficient is adaptively determined at a correlation determination 13 based on the correlation coefficient C, and the speech input signal is multiplied by the gain coefficient. The speech signal multiplied by the gain coefficient is added to an addition output Y stored in a noise memory 17 for accumulation. The accumulated signal is subtracted from the speech input signal (S+N). Since the speech signal multiplied by the gain coefficient is added to the addition output Y stored in the noise memory 17, the value of the addition output Y converges to the value of the cyclic noise N. Since the gain coefficient is adaptively determined in correspondence with the correlation coefficient C, the addition output can converge more quickly and properly than when the gain coefficient is fixed. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a signal processing circuit and a signal processing method for reducing periodic noise generated at a constant period, an imaging apparatus and an audio signal processing method for the imaging apparatus, a recording apparatus and a recording method, a reproducing apparatus, and a reproducing method.

  Imaging devices such as portable video cameras generate various noises during imaging. As an example of noise generated from the inside of the imaging apparatus, when the recording medium is a magnetic tape, for example, there is periodic noise generated at a constant period of about 150 Hz when the rotating drum rotates.

  When the imaging device is equipped with a microphone and picks up the sound and records it on a recording medium, it picks up not only the sound that is the purpose of the sound collection but also the noise generated from the inside of the imaging device, resulting in sound quality. Had fallen. Patent Document 1 describes a technique for reducing periodic noise using a synchronous adaptive filter that adaptively changes filter coefficients in accordance with periodic noise.

Japanese Patent Laid-Open No. 11-176113

  A periodic noise reduction circuit using a conventional synchronous adaptive filter will be schematically described with reference to FIG. The microphone picks up the voice input signal S + N in which the periodic noise N is mixed into the main voice signal S that is the purpose of sound collection. The audio input signal S + N is an audio signal sampled at a predetermined sampling frequency and quantized with a predetermined number of quantization bits by, for example, PCM (Pulse Code Modulation).

  The collected sound input signal S + N is input from the input terminal 100 and supplied to the addition side terminal of the adder 104. The adder 104 subtracts the filter output Y supplied from the adaptive filter 102 described later from the audio input signal S + N supplied from the input terminal 100 to obtain the audio output signal S ′. The obtained audio output signal S ′ is output from the output terminal 105.

  The audio output signal S ′ is supplied to the step gain 106. The step gain 106 supplies a residual signal E obtained by multiplying the supplied audio output signal S ′ by a predetermined gain coefficient to an LMS (Least Mean Square) calculation unit 103.

  On the other hand, a reference signal X synchronized with the period of the periodic noise N is input from the input terminal 101 and supplied to the adaptive filter 102 and the LMS operation unit 103. The reference signal X is, for example, a rotation control signal for a rotating drum supplied from a servo that controls a motor of a rotating drum (not shown), and is a signal having a period corresponding to the period of periodic noise.

Based on the reference signal X supplied from the input terminal 101 and the residual signal E supplied from the step gain 106, the LMS calculation unit 103 performs coefficient calculation so that, for example, the noise power of the residual signal E is minimized. Processing is performed, and the obtained filter coefficient W is output to the adaptive filter 102. The filter coefficient W is calculated for each sample. For example, when the number of samples is “m”, the filter coefficient W is W ₀ , W ₁ , W ₂ ,..., W _m .

The adaptive filter 102 performs adaptive processing based on the reference signal X supplied from the input terminal 101 and the filter coefficient W supplied from the LMS calculation unit 103, and the filter output Y approximated to the periodic noise N is added to the adder 104. Is output to the subtraction side terminal. This audio output signal S ′ can be expressed by Equation (1).
S ′ = S + N−Y (1)

  Since the filter output Y is a signal approximating the periodic noise N, the audio output signal S ′ becomes a signal approximating the main audio signal S by reducing the periodic noise N from the audio input signal S + N.

  Next, an example of the configuration of the adaptive filter 102 will be described with reference to FIG. The adaptive filter 102 is configured by, for example, a FIR (Finite Impulse Response) digital filter including a plurality of taps.

Reference signal X is supplied to the multiplier 111 _0, the delay elements _{110 1, 110 2, ···,} 110 m is supplied to a series-connected circuit. The delay elements 110 ₁ , 110 ₂ ,..., 110 _m output reference signals X ₁ ,..., X _m obtained by delaying the reference signal X by a sample unit, and multipliers 111 ₁ , 111 ₂ ,. -Supply to 111 _m . Also, the multipliers _{111 0,} 111 _1, 111 2, ..., the 111 _m, the filter coefficient from the LMS computation unit 103 _{W 0,} W 1, ..., _{W m} are supplied.

Multipliers _{111 0, 111 1, 111 2} , ···, 111 m , the reference signal _{X 0, X 1, ···,} X m and the filter coefficient _{W 0,} W 1, · · ·, and _{W m} Each is multiplied and output to the adder 120. The adder 120 adds the signals supplied from the multipliers 111 ₀ , 111 ₁ , 111 ₂ ,..., 111 _m and outputs a filter output Y. The filter output Y is expressed by Equation (2).

The LMS operation unit 103 calculates filter coefficients W ₀ , W ₁ ,..., W _m based on the reference signal X input to the input terminal 101 and the residual signal E supplied from the step gain 106. The filter coefficients W ₀ , W ₁ ,..., W _m are calculated using the following formula (3).
W _{k + 1} = W _k +2 μ · E _k · X _k (3)

  The value k given to the lower right of each coefficient indicates that it is the kth sample, and the value (k + 1) indicates that it is the (k + 1) th sample. “Μ” indicates a coefficient given by the step gain 106. When the value of “μ” is large, the convergence of the filter output Y output from the adaptive filter 102 can be accelerated, but the accuracy is lowered. Further, when the value of “μ” is small, the convergence of the filter output Y is delayed, while the accuracy can be increased. Therefore, an optimized value is used as the value of “μ” according to the adaptive system to be used.

  As described above, the audio output signal S ′ is fed back, and the residual signal E is used in the adaptive processing performed by the adaptive filter 102, whereby the filter coefficient W is adaptively updated, and the filter output Y is changed to the periodic noise N. Can be converged to. Then, by subtracting the filter output Y approximated to the periodic noise N from the audio input signal S + N, the periodic noise N included in the audio input signal S + N can be reduced.

  According to the noise reduction circuit using the synchronous adaptive filter as described above, in order to accelerate the convergence of the filter output Y, the value of the coefficient μ given by the step gain 106 needs to be increased. However, when the coefficient μ of the step gain 106 is increased, unnecessary signal components other than the periodic noise N are greatly reflected in the filter coefficient W, and the filter output Y that is the output of the adaptive filter based on the filter coefficient W is used. However, there is a problem in that a part of the main audio signal S, which is the purpose of sound collection, is attenuated or, conversely, a large noise is generated.

  Accordingly, an object of the present invention is to provide a signal processing circuit and a signal processing method, an image pickup apparatus, a sound signal processing method for the image pickup apparatus, and an image pickup apparatus capable of efficiently reducing periodic noise by suppressing attenuation of a main sound signal that is the purpose of sound collection Another object is to provide a recording apparatus, a recording method, a reproducing apparatus, and a reproducing method.

  In order to solve the above-described problem, the first invention is a storage unit that stores at least one period of the periodic noise period of the input audio signal, and a period of the periodic noise of the input audio signal. A correlation calculation unit for calculating a correlation between one cycle and a voice signal for one cycle one cycle before the input voice signal stored in the storage unit, and a correlation calculated by the correlation calculation unit And a correlation determination unit that determines a gain coefficient for each cycle of the cycle, accumulates the audio signal read from the storage unit by multiplying the gain coefficient for each cycle, and inputs the accumulated signal The signal processing circuit is characterized in that it is subtracted from the generated audio signal.

  Further, the second invention stores at least one period of the periodic noise period of the input audio signal in the storage unit, stores one period of the periodic noise period of the input audio signal, and the storage unit The correlation between the input speech signal stored in the input signal and the speech signal for one cycle before one cycle is calculated, and a gain coefficient is determined for each cycle of the cycle based on the calculated correlation. The signal processing method is characterized in that the audio signal read from the signal is accumulated by multiplying the gain coefficient every cycle, and the accumulated signal is subtracted from the input audio signal.

  According to a third aspect of the present invention, there is provided an image pickup unit that picks up light from a subject and outputs a video signal, a sound pickup unit built in a casing that picks up sound and outputs a sound signal, and a sound An audio signal processing unit that performs signal processing on the audio signal output from the sound collection unit, a video signal output from the imaging unit, and an audio signal output from the audio signal processing unit are recorded on a recording medium; And a recording unit having a rotation mechanism that rotates at a period of, and a sound signal processing unit stores at least one period of a periodic signal synchronized with the rotation period of the rotation mechanism of the input sound signal, and an input A correlation calculation unit for calculating a correlation between one cycle of the cycle of the periodic signal of the audio signal and the audio signal for one cycle one cycle before the input audio signal stored in the storage unit; , One cycle of the cycle based on the correlation calculated by the correlation calculation unit A correlation determination unit that determines a gain coefficient for each time, accumulates the audio signal read from the storage unit by multiplying the gain coefficient for each cycle, and stores the accumulated signal from the input audio signal The imaging apparatus is characterized in that the output is reduced.

  According to a fourth aspect of the present invention, a video signal output step of imaging light from a subject with an imaging unit and outputting a video signal, and a voice collecting unit incorporated in the housing, Output audio signal output step, audio signal processing step for performing signal processing on the audio signal output from the audio signal output step, video signal output from the video signal output step, and audio signal processing step Recording the audio signal on a recording medium with a recording unit having a rotation mechanism that rotates at a predetermined cycle, and the audio signal processing step is synchronized with at least the rotation cycle of the rotation mechanism of the input audio signal One period of the periodic signal is stored in the storage unit, one period of the period of the periodic signal of the input audio signal, and one cycle one cycle before the input audio signal stored in the storage unit And a gain coefficient is determined for each period of the period based on the calculated correlation, and the audio signal read from the storage unit is multiplied by the gain coefficient for each period. An audio signal processing method for an imaging apparatus, wherein the audio signal is stored and the accumulated signal is subtracted from the input audio signal and output.

  According to a fifth aspect of the present invention, there is provided a sound collecting unit built in the housing for collecting sound and outputting a sound signal, and a sound for performing signal processing on the sound signal output from the sound collecting unit. A signal processing unit, and a recording unit that records the audio signal output from the audio signal processing unit and has a rotation mechanism that rotates at a predetermined period, and the audio signal processing unit includes at least a rotation mechanism of the input audio signal A storage unit that stores one cycle of the periodic signal synchronized with the rotation cycle of the input signal, one cycle of the cycle of the periodic signal of the input audio signal, and the input audio signal stored in the storage unit A correlation calculation unit that calculates a correlation with an audio signal for one cycle before one cycle, and a correlation determination unit that determines a gain coefficient for each cycle of the cycle based on the correlation calculated by the correlation calculation unit; For each audio signal read from the storage unit, Accumulated by multiplying the in-factor, which is a recording apparatus which is characterized in that as subtracted from the input audio signal the stored signal.

  According to a sixth aspect of the present invention, there is provided a sound signal output step of collecting sound by a sound pickup unit built in the housing and outputting a sound signal, and a signal for the sound signal output from the sound signal output step. An audio signal processing step for performing processing, and a step of recording the audio signal output from the audio signal processing step by a recording unit having a rotation mechanism that rotates at a predetermined cycle. At least one cycle of the periodic signal synchronized with the rotation cycle of the rotation mechanism of the audio signal is stored in the storage unit, and one cycle of the cycle of the input audio signal is stored in the storage unit. The sound signal read from the storage unit is calculated by calculating a correlation with the sound signal for one cycle before the sound signal, determining a gain coefficient for each cycle based on the calculated correlation. 1 lap against An audio signal processing method of a recording apparatus characterized by accumulating multiplied by a gain coefficient, and so reduce the sound signal inputted signals accumulated for each.

  In addition, a seventh aspect of the present invention relates to a reproduction unit that has a rotation mechanism that reproduces an audio signal and rotates at a predetermined cycle, an audio signal processing unit that performs signal processing on the audio signal output from the reproduction unit, A sound output unit that outputs the sound signal output from the processing unit, and the sound signal processing unit stores at least one cycle of the periodic signal synchronized with the rotation period of the rotation mechanism of the input sound signal. And a correlation for calculating the correlation between one cycle of the cycle of the periodic signal of the input audio signal and the audio signal of one cycle one cycle before the input audio signal stored in the storage unit A calculation unit and a correlation determination unit that determines a gain coefficient for each cycle based on the correlation calculated by the correlation calculation unit, and gains for each cycle with respect to the audio signal read from the storage unit Multiply by a coefficient and accumulate, accumulated signal A reproducing apparatus characterized by subtracting from the input speech signal to output.

  According to an eighth aspect of the present invention, there is provided an audio signal reproduction step of reproducing an audio signal by a reproduction unit having a rotation mechanism that rotates at a predetermined cycle, and an audio that performs signal processing on the audio signal output from the audio signal reproduction step A signal processing step; and a step of outputting the audio signal output from the audio signal processing step, wherein the audio signal processing step corresponds to at least one period of the periodic signal synchronized with the rotation period of the rotation mechanism of the input audio signal. Is stored in the storage unit, and one cycle of the cycle of the periodic signal of the input audio signal and the audio signal of one cycle one cycle before the input audio signal stored in the storage unit A correlation is calculated, a gain coefficient is determined for each period of the period based on the calculated correlation, and the audio signal read from the storage unit is accumulated by multiplying the gain coefficient for each period. Trust A reproduction method being characterized in that so as to output the subtracted from the input speech signal.

  As described above, in the first and second inventions, at least one period of the periodic noise period of the input audio signal is stored in the storage unit, and 1 period of the periodic noise period of the input audio signal is stored. The correlation between the period and the input audio signal stored in the storage unit and the audio signal for one period before one period is calculated, and the gain coefficient for each period of the period based on the calculated correlation The sound signal read from the storage unit is multiplied by a gain coefficient for each cycle and accumulated, and the accumulated signal is subtracted from the input sound signal. Periodic noise can be reduced from the signal.

  In the third and fourth aspects of the invention, the light from the subject is imaged and a video signal is output, and the audio signal output by collecting the sound is subjected to signal processing. Is recorded on a recording medium by a recording unit having a rotation mechanism that rotates at a predetermined cycle, and at least one period of a periodic signal synchronized with the rotation period of the rotation mechanism of the input audio signal is stored in the storage unit. And calculating a correlation between one cycle of the cycle of the periodic signal of the input audio signal and the audio signal of one cycle one cycle before the input audio signal stored in the storage unit, Based on the correlation, a gain coefficient is determined for each cycle, the audio signal read from the storage unit is multiplied by the gain coefficient for each cycle, and the accumulated signal is stored from the input audio signal. Since the output is reduced, the original voice input Degree from, generated by the rotating mechanism the periodic noise can be recorded audio signal is reduced.

  In the fifth and sixth aspects of the invention, signal processing is performed on an audio signal output by collecting audio, and the audio signal is recorded by a recording unit having a rotation mechanism that rotates at a predetermined cycle. One cycle of the periodic signal synchronized with the rotation cycle of the rotation mechanism of the input audio signal is stored in the storage unit, and one cycle of the cycle of the input audio signal is stored in the storage unit. The correlation between the input audio signal and the audio signal for one cycle before one cycle is calculated, the gain coefficient is determined for each cycle of the cycle based on the correlation, and read from the storage unit Since the audio signal is accumulated by multiplying the gain coefficient for each cycle, and the accumulated signal is subtracted from the inputted audio signal, the periodicity generated by the rotating mechanism from the original audio input signal. Record audio signals with reduced noise It is possible.

  In the seventh and eighth inventions, signal processing is performed on an audio signal reproduced by a reproduction unit having a rotation mechanism that rotates at a predetermined cycle, and an audio signal is output and input. At least one cycle of the periodic signal synchronized with the rotation cycle of the rotation mechanism of the audio signal is stored in the storage unit, and one cycle of the cycle of the input audio signal is stored in the storage unit. The correlation between the audio signal and the audio signal for one cycle one cycle before is calculated, a gain coefficient is determined for each cycle based on the correlation, and the audio signal read from the storage unit is calculated. Since the gain signal is multiplied every cycle and accumulated, and the accumulated signal is subtracted from the input audio signal for output, the audio in which periodic noise is mixed with the main audio input signal during recording Reduces periodic noise from the signal It is possible to reproduce the audio signal.

  The present invention sets a gain based on the correlation between an audio input signal mixed with periodic noise and the audio input signal of the previous cycle, generates a signal by accumulating periodic noise components, and the accumulated signal. Is subtracted from the voice input signal, so that the periodic noise can be reduced from the original voice input signal.

  Further, according to the present invention, when the correlation between the voice input signal and the voice input signal of the previous cycle is low, the gain signal is reduced and the output signal is accumulated. Accumulation of signal components can be minimized.

  The first embodiment of the present invention will be described below. The present invention adaptively sets a gain coefficient based on a correlation between a voice input signal in which periodic noise generated at a fixed period is mixed in a main voice signal, which is the purpose of sound collection, and a voice input signal one cycle before. The sound input signal is multiplied by the gain coefficient and accumulated, and the accumulated signal is subtracted from the sound input signal.

  FIG. 1 shows an example of the configuration of a periodic noise reduction circuit 1 according to the first embodiment of the present invention. For example, when the main sound signal S, which is the purpose of sound collection, is picked up by a microphone, the periodic noise N derived from mechanical sound generated from the device is actually collected together with the main sound signal S. A sound input signal S + N in which periodic noise N is mixed with the main sound signal S is input to the input terminal 10. The audio input signal S + N, which is the input of the main audio signal S and the periodic noise N and is input to the input terminal 10, is supplied to the buffer memory 11, the correlation calculation unit 12, and the addition side terminal of the adder 18.

  The audio input signal S + N is a digital audio signal sampled by a PCM at a predetermined sampling frequency and quantized with a predetermined number of quantization bits. For example, when the periodic noise reduction circuit 1 is used in a video camera using a magnetic tape as a recording medium, the periodic noise N is noise generated when the rotating drum rotates at a constant period. In another example, when the noise reduction circuit 1 is used in a video camera using a disk-shaped recording medium such as an optical disk as a recording medium, the noise is generated when the spindle motor rotates at a constant period.

  On the other hand, a reference signal X synchronized with the period of the periodic noise N is input to the input terminal 19 and supplied to the address generator 20 and the correlation calculator 12. The reference signal X is, for example, a pulse signal output in synchronization with the period of the periodic noise N. For example, when the periodic noise reduction circuit 1 is used in a video camera using a magnetic tape as a recording medium, the reference signal X is a rotation drum rotation control signal obtained from a servo that controls the rotation drum motor. is there. In another example, the reference signal X is a rotation obtained from a spindle motor driver that controls the spindle motor when the noise reduction circuit 1 is used in a video camera using a disk-shaped recording medium such as an optical disk as a recording medium. It is a control signal.

  The address generation unit 20 detects the rising or falling edge of the reference signal X supplied from the input terminal 19, and generates an address incremented by one, for example, in units of samples based on the detection result. The generated address is supplied to the buffer memory 11 and the noise memory 17. The read position and / or write position of the buffer memory 11 and the noise memory 17 are specified by this address.

  The buffer memory 11 has a capacity capable of storing an audio signal for at least one period of the periodic noise N of the audio input signal S + N. For example, the buffer memory 11 can store an audio input signal S + N for one sample at one address. Based on the address supplied from the address generator 20, the buffer memory 11 can read out the voice input signal one cycle before the periodic noise N in units of samples. The audio input signal read from the buffer memory 11 is supplied to the correlation calculation unit 12 and the addition side terminal of the adder 14. Also, based on the address supplied from the address generation unit 20, the audio input signal S + N input to the input terminal 10 is written in the buffer memory 11 in units of samples.

  The correlation calculation unit 12 is for one period of the periodic noise N between the current voice input signal S + N supplied from the input terminal 10 and the voice input signal one cycle before the periodic noise N supplied from the buffer memory 11. The correlation coefficient C which shows the correlation for 1 period of the periodic noise N of two audio | voice input signals is calculated. The correlation coefficient C is calculated by, for example, equation (4).

“X _i ” indicates each sample value of the audio input signal S + N, and “X bar” indicates an average value of each sample value for one period of the periodic noise N of the audio input signal S + N. “Y _i ” represents each sample value of the audio input signal supplied from the buffer memory 11, and “Y bar” represents each sample value for one cycle of the audio input signal supplied from the buffer memory 11. Average values are shown. “N” indicates the number of samples in one period. The values of “X bar” and “Y bar” are values obtained by averaging the sample values for one period of each audio input signal. In the case of an audio signal, the value of the DC component is “0” and the sum of the values of the other AC components is considered to be “0”. The number C can be calculated. The value of the correlation coefficient C ranges from “+1” to “−1”. The closer the value is to “+1”, the more the correlation between the audio input signal and the audio input signal supplied from the buffer memory 11 is. Is high. When the value of the correlation coefficient C is “−1”, it indicates that one audio signal and the other audio signal are in an opposite phase relationship. It is preferred to use values from 0 "to" +1 ".

  The calculated correlation coefficient C is supplied to the correlation determination unit 13 in synchronization with the rising or falling timing of the reference signal X. The correlation determination unit 13 sets a gain coefficient according to the correlation coefficient C supplied from the correlation calculation unit 12. The set gain coefficient is supplied to the multiplier 15.

  For example, the gain coefficient is set as follows. If the value of the correlation coefficient C is large and the correlation between the voice input signal S + N and the voice input signal one cycle before the periodic noise N output from the buffer memory 11 is high, it is included in the voice input signal S + N. In the audio signal, since the periodic noise N is considered to be dominant, the gain coefficient is set to a large value. On the other hand, when the value of the correlation coefficient C is small and the correlation between the voice input signal S + N and the voice input signal one cycle before the periodic noise N output from the buffer memory 11 is low, the voice input signal S + N Since the included audio signal is considered to be dominant audio signals other than the periodic noise N, the gain coefficient is set to a small value.

  That is, when the correlation between the voice input signal S + N and the voice input signal one cycle before the periodic noise N is high, the voice input signal S + N is considered to have the periodic noise N dominant over the main voice signal S. Therefore, by increasing the gain coefficient, more periodic noise components are accumulated. On the other hand, when the correlation is low, the main audio signal S is considered to be dominant over the periodic noise N. Therefore, by reducing the gain coefficient, signal components other than the periodic noise N are accumulated. Try to suppress.

  As a more specific setting method of the gain coefficient, for example, a threshold is provided for the value of the correlation coefficient C, the value of the correlation coefficient C is compared with the threshold, and a fixed value is determined based on the comparison result. It is possible to use it. When the value of the correlation coefficient C exceeds the threshold, the gain coefficient is set to a large predetermined value, for example, “1/64”, and when the value of the correlation coefficient C is equal to or smaller than the threshold, the gain coefficient is set to a small predetermined value, For example, “1/256” is set. The set gain coefficient is supplied to the multiplier 15.

  Note that the above-described gain coefficient and threshold value are merely examples, and the present invention is not limited thereto, and other values may be used. Further, the method for setting the gain coefficient is not limited to this example. For example, two or more threshold values may be provided for the value of the correlation coefficient C, and the gain coefficient may be set stepwise according to the threshold value. Further, for example, a continuous value may be used as the gain coefficient with respect to the value of the correlation coefficient C.

  The audio signal read from the buffer memory 11 is supplied to the addition side terminal of the adder 14. On the other hand, a signal supplied from a noise memory 17 described later is supplied to the subtraction side terminal of the adder 14. The adder 14 subtracts the signal supplied from the noise memory 17 from the audio signal supplied from the buffer memory 11 and supplies the result to the multiplier 15. The multiplier 15 multiplies the signal supplied from the adder 14 by the gain coefficient supplied from the correlation determination unit 13 and outputs the result to the adder 16. The adder 16 adds the signal supplied from the multiplier 15 and the signal supplied from the noise memory 17. The addition output Y from the adder 16 is supplied to the noise memory 17 and also supplied to the subtraction side terminal of the adder 18.

  The noise memory 17 has a capacity capable of storing an audio signal for one period of the periodic noise N, and can store an audio signal for one sample at one address, for example. The noise memory 17 is configured to read the stored audio signal in units of samples based on the address supplied from the address generation unit 20. That is, the addition output Y one cycle before the periodic noise N is read out. Further, based on the address supplied from the address generation unit 20, the addition output Y supplied from the adder 16 is written in the noise memory 17 in units of samples.

  The adder 18 subtracts the addition output Y of the adder 16 from the audio input signal S + N supplied from the input terminal 10 to obtain an audio output signal S ′. The obtained audio output signal S ′ is output from the output terminal 21. This audio output signal S ′ is a signal obtained by reducing the component of the periodic noise N from the audio input signal S + N.

  Next, the operation of the periodic noise reduction circuit 1 according to the first embodiment of the present invention will be described with reference to FIG. FIG. 2 shows an example of the signal waveform of each part in the periodic noise reduction circuit 1 described above. Here, the signal is divided into sections A, B and C for each period of the reference signal X, and the signal waveforms in the respective sections will be described.

  Here, as a method of setting the gain coefficient in the correlation determination unit 13, a threshold value is set in advance for the correlation coefficient C, and when the value of the correlation coefficient C is equal to or larger than this threshold value and when it is equal to or smaller than the threshold value. On the other hand, a case where a fixed gain coefficient is set for each will be described. As an example, the threshold is set to “0.6”, the gain coefficient when the correlation coefficient C is equal to or greater than the threshold is set to “1/64”, and the gain coefficient when the correlation coefficient C is equal to or less than the threshold is “ 1/256 ".

  In the section A, the correlation coefficient C indicating the correlation between the voice input signal S + N shown in FIG. 2A and the voice input signal one cycle before the periodic noise N output from the buffer memory 11 shown in FIG. 12 is calculated. The correlation coefficient C is output from the correlation calculation unit 12 at the head of the section B as shown in FIG. 2E in synchronization with the rising edge of the reference signal X shown in FIG. 2B. The correlation determination unit 13 sets the gain coefficient shown in FIG. 2F based on the correlation coefficient C. The gain coefficient is supplied to the multiplier 15.

  In this example, the audio input signal (FIG. 2C) one cycle before the periodic noise N output from the buffer memory 11 in the section A is other than the periodic noise compared to the input audio input signal S + N (FIG. 2A). Is a signal on which the audio signal is slightly superimposed. That is, it is considered that the voice input signal S + N and the voice input signal output from the buffer memory 11 have a slightly low correlation. Here, for example, “0.5” is calculated as the correlation coefficient C indicating the correlation between the voice input signal S + N in the section A and the voice input signal one cycle before the periodic noise N output from the buffer memory 11. It shall be assumed. The correlation determination unit 13 compares the correlation coefficient C (= 0.5) supplied from the correlation calculation unit 12 with a preset threshold value (= 0.6). As a result of the comparison, it is determined that the correlation coefficient C is equal to or less than the threshold value, and a gain coefficient “1/256” is output.

  The audio signal output from the noise memory 17 illustrated in FIG. 2D is subtracted in the section B from the audio signal output from the buffer memory 11 illustrated in FIG. 2C, and the gain coefficient output from the correlation determination unit 13 to the subtracted audio signal. By multiplying “1/256”, an audio signal output from the multiplier 15 shown in the section B of FIG. 2G is obtained. This audio signal and the audio signal output from the noise memory 17 shown in the section B of FIG. 2D are added by the adder 16 to obtain an addition output Y of the adder 16 shown in FIG. 2H. The addition output Y is subtracted from the audio input signal S + N input to the input terminal 10 by the adder 18, and is output from the output terminal 21 as the audio output signal S 'shown in FIG. 2I. Further, the addition output Y is input to and stored in the noise memory 17.

  Further, in the section B, the correlation coefficient C indicating the correlation between the speech input signal S + N (FIG. 2A) and the speech input signal one cycle before the periodic noise N output from the buffer memory 11 (FIG. 2C) is a correlation calculation. Calculated by the unit 12. The correlation coefficient C is output from the correlation calculation unit 12 at the head of the section C as shown in FIG. 2E in synchronization with the reference signal X (FIG. 2B). The correlation determination unit 13 sets a gain coefficient (FIG. 2F) based on the correlation coefficient C. The gain coefficient is supplied to the multiplier 15.

  In this example, the audio input signal (FIG. 2C) one period before the periodic noise N output from the buffer memory 11 in the section B is other than the periodic noise compared to the input audio input signal S + N (FIG. 2A). The audio signal is hardly superimposed. That is, the audio input signal S + N and the audio input signal output from the buffer memory 11 are considered to have a high correlation. Here, for example, “0.9” is calculated as the correlation coefficient C indicating the correlation between the voice input signal S + N in the section B and the voice input signal one cycle before the periodic noise N output from the buffer memory 11. It shall be assumed. For example, the correlation determination unit 13 compares the correlation coefficient C (0.9) supplied from the correlation calculation unit 12 with a preset threshold value (0.6). It is determined that the value is equal to or greater than the threshold, and a gain coefficient “1/64” is output.

  The audio signal (FIG. 2D) output from the noise memory 17 is subtracted in the section C from the audio signal (FIG. 2C) output from the buffer memory 11, and the gain coefficient output from the correlation determination unit 13 to the subtracted audio signal. By multiplying by “1/64”, an audio signal output from the multiplier 15 shown in section C of FIG. 2G is obtained. This audio signal and the audio signal output from the noise memory 17 shown in section C of FIG. 2D are added by the adder 16 to obtain an added output Y (FIG. 2H). The addition output Y is subtracted from the audio input signal S + N input to the input terminal 10 by the adder 18 and is output from the output terminal 21 as the audio output signal S ′ (FIG. 2I). Further, the addition output Y is input to and stored in the noise memory 17.

  In this way, by adding the audio signal multiplied by the gain coefficient to the addition output Y stored in the noise memory 17, the value of the addition output Y converges to the value of the periodic noise N. At this time, since the gain coefficient is adaptively set according to the correlation coefficient C, it is possible to converge more quickly and appropriately than when the gain coefficient is fixed.

  The portion composed of the adder 14, the multiplier 15, the adder 16 and the noise memory 17 is a low-pass filter having the output of the buffer memory 11 as an input, and has a fixed phase with periodic noise N. Since the frequency component is substantially direct current, periodic noise N can be extracted by applying a low-pass filter to each phase of the main audio signal S.

  If the period of the periodic noise is an integer multiple of the period of the reference signal X, the periodic noise can be reduced.

  Next, a second embodiment of the present invention will be described with reference to FIG. In the second embodiment of the present invention, the periodic noise reduction circuit 1 according to the first embodiment of the present invention is provided with a limiter 30 for limiting the level of the audio input signal S + N for performing the correlation calculation. . When the level of the audio input signal S + N is limited, it is assumed that the correlation with the audio input signal S + N of the previous cycle is low, and the gain coefficient is set to a fixed value according to the occurrence of the level limitation. . In the following description, portions common to those in FIG. 1 described above are denoted by the same reference numerals and description thereof is omitted.

  The limiter 30 limits the level of the audio input signal S + N supplied from the input terminal 10 based on a positive threshold value + TH and a negative threshold value -TH set in advance. The threshold values + TH and -TH are values that, for example, measure the maximum level of the periodic noise N in advance and do not limit the periodic noise component. As an example, when the level of the periodic noise N can be expressed by 4 bits, the threshold values + TH and -TH are values that limit the level for a signal exceeding 4 bits. When the level of the voice input signal S + N supplied from the input terminal 10 exceeds the threshold value + TH or −TH, the limiter 30 limits the level of the voice input signal S + N to the threshold value, and the correlation calculation unit 12 and the buffer memory 11. Output for.

  The correlation calculation unit 12 performs a correlation calculation for one cycle of the periodic noise N between the voice input signal supplied from the limiter 30 and the voice input signal one cycle before the periodic noise N supplied from the buffer memory 11. A correlation coefficient C indicating the correlation between two audio signals is calculated in the same manner as in the first embodiment. The calculated correlation coefficient C is output to the correlation determination unit 13 in synchronization with the reference signal X. Here, the correlation calculation unit 12 determines whether or not the audio input signal supplied from the limiter 30 and the audio input signal one cycle before the periodic noise N supplied from the buffer memory 11 are limited by the limiter 30. Determine whether.

  As a result of the determination, for example, at least one of the audio input signal supplied from the limiter 30 and the audio input signal one cycle before the periodic noise N supplied from the buffer memory 11 is transmitted by the limiter 30. In the case of the limited signal, it can be determined that the correlation between the audio input signal supplied from the limiter 30 and the audio input signal one cycle before the periodic noise N supplied from the buffer memory 11 is low. Therefore, the correlation calculation unit 12 uses the correlation coefficient C to determine the correlation between the voice input signal supplied from the limiter 30 and the voice input signal one cycle before the periodic noise N supplied from the buffer memory 11. A value indicating low, for example, “0” is set.

  The determination as to whether or not the signal is limited by the threshold value may be performed, for example, by detecting a sample having the same value as the threshold value when the correlation calculation unit 12 calculates the correlation coefficient C. For example, when even one sample having the same value as the threshold is detected, the value of the correlation coefficient C is set to one cycle of the audio input signal supplied from the limiter 30 and the periodic noise N supplied from the buffer memory 11. A value indicating that the correlation with the previous voice input signal is low, for example, “0” may be considered. Further, for example, when the number of samples having the same value as the threshold is counted and the number of samples is equal to or larger than a predetermined number, the value of the correlation coefficient C is changed to the audio input signal supplied from the limiter 30 and the periodicity supplied from the buffer memory 11. It is conceivable to set a value indicating that the correlation between the noise N and the voice input signal one cycle before is low, for example, “0”.

  Further, for example, by supplying a control signal indicating that the level of the audio input signal is limited by the limiter 30 from the limiter 30 to the correlation calculation unit 12, it is determined whether the signal is limited by the threshold value. It can also be done.

  The correlation determination unit 13 sets a gain coefficient according to the correlation coefficient C supplied from the correlation calculation unit 12. A value indicating that the correlation coefficient C is low in correlation between the voice input signal supplied from the limiter 30 and the voice input signal one cycle before the periodic noise N supplied from the buffer memory 11, for example, “0 ”, It is determined that the correlation between the audio input signal supplied from the limiter 30 and the audio input signal one cycle before the periodic noise N supplied from the buffer memory 11 is low. The unit 13 sets the gain coefficient to “0”, for example.

  Next, the operation of the periodic noise reduction circuit 1 'according to the second embodiment of the present invention will be described with reference to FIG. FIG. 4 shows an example of the signal waveform of each part in the periodic noise reduction circuit 1 'according to the second embodiment of the present invention. Here, a value indicating that the correlation between the audio input signal supplied from the limiter 30 and the audio input signal one cycle before the periodic noise N supplied from the buffer memory 11 is low is described as “0”. To do.

  Threshold values + TH and -TH are set in advance in the limiter 30, and in the section A, a part of the audio input signal S + N shown in FIG. 4A input to the limiter 30 exceeds the threshold values + TH and -TH. As shown in 4C, a part of the audio input signal S + N is limited to a threshold value and output. The part indicated by “K” in FIG. 4C indicates that the voice input signal S + N is restricted by the threshold value + TH, and the part indicated by “L” indicates that the voice input signal S + N is restricted by the threshold value −TH. Show.

  In the correlation calculation unit 12, a correlation coefficient C indicating a correlation between the voice input signal (FIG. 4C) whose level is limited by the limiter 30 and the voice signal output from the buffer memory 11 shown in FIG. Is calculated by The correlation coefficient C is supplied from the correlation calculation unit 12 to the correlation determination unit 13 at the head of the section B as shown in FIG. 4F in synchronization with the reference signal X shown in FIG. 4B.

  In this example, in the section A, the audio input signal (FIG. 4C) output from the limiter 30 is limited by the threshold values + TH and −TH. Therefore, “0” is output as the value of the correlation coefficient C. When the value of the correlation coefficient C supplied from the correlation calculation unit 12 is “0”, the correlation determination unit 13 calculates the difference between the audio input signal supplied from the limiter 30 and the audio signal supplied from the buffer memory 11. It is determined that the correlation is low, and “0” that is a gain coefficient when the correlation is low is output.

  Furthermore, in the section B, the audio input signal S + N (FIG. 4A) input to the limiter 30 is not a signal exceeding the threshold values + TH and −TH set in the limiter 30, and thus is output without being level limited. The correlation calculation unit 12 calculates a correlation coefficient C indicating the correlation between the voice input signal (FIG. 4C) supplied from the limiter 30 and the voice signal (FIG. 4D) output from the buffer memory 11. The correlation coefficient C is supplied from the correlation calculation unit 12 to the correlation determination unit 13 at the head of the interval C as shown in FIG. 4F in synchronization with the reference signal X.

  In this example, the audio signal (FIG. 4D) output from the buffer memory 11 in the section B is limited by the threshold values + TH and −TH (parts “K ′” and “L ′” in the figure). . Therefore, “0” is output as the value of the correlation coefficient C. When the value of the correlation coefficient C supplied from the correlation calculation unit 12 is “0”, the correlation determination unit 13 calculates the difference between the audio input signal supplied from the limiter 30 and the audio signal supplied from the buffer memory 11. It is determined that the correlation is low, and “0” is output as the gain coefficient when the correlation is low. In the sections B and C, since the gain coefficients output from the correlation determination unit 13 are both “0”, the output of the multiplier 15 shown in FIG. 4H is “0”, and the noise memory 17 shown in FIG. The signal output from is stored in the noise memory 17 as it is.

  The audio input signal S + N that exceeds the maximum level of the periodic noise N is considered to be dominant in the main audio signal S. In the second embodiment of the present invention, when the main audio signal S is dominant in the audio input signal S + N in this way, the gain coefficient is set to “0” and no signal components other than the periodic noise N are accumulated. Like that. For this reason, it is possible to prevent the main audio signal component from affecting the added output Y and efficiently converge the added output Y to the periodic noise N.

  Also, by limiting the level of the audio input signal S + N by the limiter 30, the number of bits of the audio signal stored in the buffer memory 11 and the noise memory 17 can be reduced to the number of bits that can represent periodic noise, and the buffer The capacity of the memory 11 and the noise memory 17 can be reduced.

  Next explained is the third embodiment of the invention. In the third embodiment of the present invention, an initial value of periodic noise is detected, and the initial value is written in the noise memory as initial data of the noise memory. FIG. 5 shows a configuration of an example of a periodic noise reduction circuit 1 ″ according to the third embodiment of the present invention. In the following, the same reference numerals are used for parts common to FIGS. 1 and 3 described above. The description is omitted.

  The periodic noise reduction circuit 1 ″ according to the third embodiment of the present invention is different from the periodic noise reduction circuit 1 ′ according to the second embodiment described above in that the selector 31, the selector 32, and the initial noise capturing determination. The selector 31 is provided between the limiter 30 and the buffer memory 11, and switches the output of the limiter 30 and the output of the buffer memory 11 to be supplied to the buffer memory 11. The limiter 30 Is supplied to the correlation calculation unit 12 via the selector 31. The selector 32 is provided between the adder 16 and the noise memory 17, and the output of the adder 16 is output from the adder 16. The output of the predetermined value and the output of the buffer memory 11 are switched and output to the noise memory 17 and the adder 18.

  The initial noise capture determination unit 33 receives the reference signal X input to the input terminal 19 and the correlation coefficient C output from the correlation calculation unit 12, and uses the reset signal R input from the input terminal 34 as a trigger. The initial noise capturing operation is started based on the correlation coefficient C supplied from the correlation calculation unit 12. The initial noise capturing determination unit 33 controls switching between the input terminal of the selector 31 and the input terminal of the selector 32 in accordance with the initial noise capturing operation.

  Next, the operation of the periodic noise reduction circuit 1 ″ according to the third embodiment of the present invention will be described with reference to FIG. 6. FIG. An example of the signal waveform of each part is shown. Note that the value of the correlation coefficient C is always a high value, for example, “0.9”. Here, the operation for setting the initial value of the output signal will be described, and the operation after setting the initial value of the output signal (after the section F) is the same as that in the second embodiment described above. Is omitted.

  FIG. 6E shows the selection state of the input terminal of the selector 31. In this example, the symbols “p” and “q” indicate that the input terminals 31a and 31b are selected, respectively. FIG. 6F shows the selection state of the input terminal of the selector 32. In this example, the symbols “r”, “s”, and “t” indicate that the input terminals 32a, 32b, and 32c are selected, respectively.

  In section A, which is the initial state, the selector 31 selects the input terminal 31b, and the selector 32 selects the input terminal 32b. The audio input signal S + N shown in FIG. 6B is supplied to the adder 18 and is also supplied to the selector 31 and the correlation calculation unit 12 via the limiter 30. In the selector 31, since the input terminal 31 b is selected, the audio input signal supplied from the limiter 30 is supplied to the buffer memory 11 and stored therein.

  A reset signal R shown in FIG. 6A is generated at a predetermined timing such as when the apparatus is turned on. The reset signal R is input from the input terminal 34 and supplied to the initial noise capturing determination unit 33. For example, the initial noise capture determination unit 33 uses the rising edge of the reset signal R as a trigger to synchronize with the reference signal X shown in FIG. 6C and sets the input terminal of the selector 31 at the head of the next section B where the reset signal R rises. The selector 31 is controlled to switch to 31a.

  In the section B, the audio signal (FIG. 6D) output from the buffer memory 11 is supplied to the selector 31 and the correlation calculation unit 12. As described above, the selector 31 selects the input terminal 31a under the control of the initial noise acquisition determination unit 33, so that the audio signal output from the buffer memory 11 is supplied to the correlation calculation unit 12, The data is supplied to and stored in the buffer memory 11 as it is, and the audio signal in the buffer memory 11 is held. The correlation calculation unit 12 calculates a correlation coefficient C from the audio signal supplied from the buffer memory 11 shown in FIG. 6D and the audio input signal supplied from the limiter 30, and the initial noise capturing determination unit 33 and the correlation determination unit 13 is supplied.

  The initial noise capture determination unit 33 monitors the correlation coefficient C, and detects a correlation coefficient C indicating that the correlation between the audio signal supplied from the buffer memory 11 and the audio input signal supplied from the limiter 30 is high. The selected number of times is counted for each period, and the selection state of the input terminals of the selector 31 and the selector 32 is held until the correlation coefficient C indicating that the correlation is high is continuously detected for a predetermined period.

  In this example, the selector 31 until the correlation coefficient C indicating that the correlation between the audio signal supplied from the buffer memory 11 and the audio input signal supplied from the limiter 30 is high is detected for three consecutive cycles. The selection state of the input terminal of the selector 32 is held. In the selector 31, the state in which the input terminal 31 a is selected is maintained from the section B to the section D. Similarly, in the selector 32, the state in which the input terminal 32b is selected from the section B to the section D is held.

  Since the selection state of the input terminal of the selector 31 is held in the sections C and D, the audio signal (FIG. 6D) output from the buffer memory 11 is input to and held in the buffer memory 11, and The held audio signal of the buffer memory 11 is supplied to the correlation calculation unit 12. The correlation calculation unit 12 calculates a correlation coefficient C from the audio signal supplied from the buffer memory 11 and the audio input signal supplied from the limiter 30, and supplies the correlation coefficient C to the initial noise capturing determination unit 33 and the correlation determination unit 13. .

  As a result of monitoring the correlation coefficient C, the initial noise capture determination unit 33 has a correlation coefficient C indicating that the correlation between the audio signal supplied from the buffer memory 11 and the audio input signal supplied from the limiter 30 is high. Since it was detected for three consecutive periods from section B (between section B and section D), the selector 31 is controlled to switch the input terminal of the selector 31 to 31b at the beginning of the section E in synchronization with the reference signal X. The selector 32 is controlled to switch the input terminal of the selector 32 to 32c.

  In the section E, since the input terminal 32 c is selected in the section E, the audio signal (FIG. 6D) output from the buffer memory 11 is supplied to the noise memory 17 and stored as initial data of the noise memory 17. . Since the input terminal 31b is selected in the selector 31, the audio input signal output from the limiter 30 is supplied to the buffer memory 11, and the audio signal in the buffer memory 11 is updated. The audio input signal output from the limiter 30 is also supplied to the correlation calculation unit 12. Further, the initial noise acquisition determination unit 33 controls the selector 32 so as to switch the input terminal of the selector 32 to 32a at the head of the section F in synchronization with the reference signal X.

  After the section F, the periodic noise reduction operation is performed in the same manner as in the second embodiment described above. The correlation calculation unit 12 calculates a correlation coefficient C from the audio signal supplied from the buffer memory 11 and the audio input signal supplied from the limiter 30, and based on the correlation coefficient C, the correlation determination unit 13 A gain coefficient is set and supplied to the multiplier 15. The adder 14 subtracts the audio signal output from the noise memory 17 from the audio signal output from the buffer memory 11 and supplies the result to the multiplier 15. The multiplier 15 multiplies the audio signal supplied from the adder 14 by the gain coefficient supplied from the correlation determination unit 13 and outputs the result to the adder 16. The adder 16 adds the audio signal output from the noise memory 17 to the signal supplied from the multiplier 15, and supplies the addition output Y shown in FIG. 6G to the adder 18 and the noise memory 17. The adder 18 subtracts the addition output Y from the audio input signal S + N to obtain the audio output signal S ′ shown in FIG. 6H.

  Note that the initial noise capture determination unit 33 has a correlation coefficient C indicating that the correlation between the audio signal supplied from the buffer memory 11 and the audio input signal supplied from the limiter 30 is high after the trigger by the reset signal R. If it is not detected continuously for a predetermined period, for example, it may reset itself and perform the initial noise capturing operation shown in FIG. 6 again. At this time, the initial noise acquisition determination unit 33 controls to select the input terminal 32 b of the selector 32 and clears the audio signal stored in the noise memory 17. Further, for example, the value of the initial data in the noise memory 17 can be set to “0” to perform the periodic noise reduction operation as shown in the second embodiment.

  The operation using the limiter 30 can be performed in the same manner as in the second embodiment described above. Further, the limiter 30 can be omitted in the third embodiment.

  When the main audio signal S is dominant among the signal components of the audio input signal S + N, it is generally considered that the correlation with the audio input signal in another period is low. On the other hand, when the periodic noise N is dominant, it is considered that the correlation with the voice input signal containing almost no signal components other than the periodic noise N in another period is high. Therefore, when the correlation coefficient C indicating that the correlation is high is detected continuously for a predetermined period, it is considered that signal components other than the periodic noise N are hardly included, and this audio input signal is stored in the noise memory 17. Thus, the convergence of the addition output Y to the periodic noise N can be accelerated.

  Next explained is the fourth embodiment of the invention. 7 shows a configuration of an example of an imaging device 50 to which the periodic noise reduction circuit 1, 1 ′, or 1 ″ according to the first, second, or third embodiment of the present invention is applied. As a recording medium, for example, a disk-shaped recording medium is used, and periodic noise caused by a spindle motor for rotating the disk-shaped recording medium, which is generated during recording, is efficiently reduced.

  As the recording medium, for example, a recordable type DVD (Digital Versatile Disc) can be used as the disc-shaped recording medium. For example, a recordable type CD (Compact Disc) may be used as the disc-shaped recording medium. Further, for example, a hard disk or a magneto-optical disk may be used as the recording medium. In the following, it is assumed that the recording medium is a recordable DVD.

  The imaging device 50 includes a camera unit 51, a recording / playback processing unit 52, and a control unit 53. The camera unit 51 includes an optical block 60, a camera control unit 61, a signal converter 62, an imaging signal processing unit 63, a built-in microphone 64, and an audio signal processing unit 65. The optical block 60 includes a lens group for imaging a subject, an aperture adjustment mechanism, a focus adjustment mechanism, a zoom mechanism, a shutter mechanism, a flash mechanism, a camera shake correction mechanism, and the like. The camera control unit 61 receives a control signal from the control unit 53 and generates a control signal to be supplied to the optical block 60. Then, the generated control signal is supplied to the optical block 60 to perform controls such as zoom control, shutter control, and exposure control.

  The signal converter 62 is configured by an image sensor such as a CCD (Charge Coupled Device), for example, and an image through the optical block 60 is formed on an image forming surface thereof. The signal converter 62 receives a video capturing timing signal supplied from the control unit 53 according to the shutter operation, converts the subject image formed on the imaging plane into an imaging signal, and the imaging signal processing unit 63. To supply.

  The imaging signal processing unit 63 performs processing such as gamma correction and AGC (Auto Gain Control) on the imaging signal based on the control signal from the control unit 53, and converts the imaging signal into a video signal as a digital signal. I do. Further, the imaging signal processing unit 63 further performs automatic white balance control, exposure correction control, and the like on the video signal based on the control signal from the control unit 53.

  The built-in microphone 64 is built in the housing of the imaging device 50, converts the collected sound into an electrical signal, and outputs it as a sound signal. The built-in microphone 64 collects, for example, periodic noise generated from a spindle motor, for example, together with the main sound that is the purpose of sound collection. The audio signal from the built-in microphone 64 is supplied to the audio signal processing unit 65. The audio signal processing unit 65 performs processing such as sound quality correction and AGC on the audio signal supplied from the built-in microphone 64 based on the control signal from the control unit 53, and converts the audio signal into a digital audio signal. .

  The audio signal processing unit 65 includes the periodic noise reduction circuit 1, 1 ′, or 1 ″ described in the first, second, or third embodiment described above. Then, periodic noise reduction processing is performed.

  FIG. 8 shows an exemplary configuration of the audio signal processing unit 65. The audio signal processing unit 65 includes, for example, an analog processing unit 94, an A / D conversion unit 95, and a periodic noise reduction unit 96. The audio signal supplied from the built-in microphone 64 is subjected to processing such as sound quality correction and AGC as required by the analog processing unit 94, and converted into a digital audio signal by the A / D conversion unit 95. That is, the digital audio signal is the digital audio input signal S + N in which the periodic noise N generated from the spindle motor is mixed with the main audio signal S described in the first embodiment. The audio input signal S + N is supplied to the periodic noise reduction unit 96.

  The periodic noise reduction unit 96 uses a reference signal X supplied from a spindle motor driver that controls a spindle motor of the disk drive 200 described later, to the audio input signal S + N supplied from the A / D conversion unit 95. Then, the periodic noise reduction processing is performed by the method described in the first, second, or third embodiment, and the audio output signal S ′ is obtained by removing the periodic noise from the audio input signal S + N. The obtained audio signal S ′ is supplied to the recording / reproducing processor 52.

  Returning to FIG. 7, the recording / playback processing unit 52 includes an encoding / decoding unit 70, a buffer memory 71, and an output processing unit 72. The encoding / decoding unit 70 compresses and multiplexes the video signal, audio signal, and additional recording information from the camera unit 51 by using a buffer memory 71 made of, for example, SDRAM (Synchronous Dynamic Random Access Memory). In addition, the encoding / decoding unit 70 separates the video signal, the audio signal, and the additional recording information from the compressed and multiplexed data using the buffer memory 71, and decodes them.

  The output processing unit 72 supplies the compressed data from the encoding / decoding unit 70 to the control unit 53 and the output terminals 73 to 75 under the control of the control unit 53.

  The control unit 53 includes a system controller 80, a ROM (Read Only Memory) 81, a RAM (Random Access Memory) 82, an operation input interface 83 for connecting the operation input unit 90, and a display control unit for connecting the display unit 91. 84, a memory card interface 85 for loading a memory card 92, an angular velocity detector 86 for detecting angular velocity for camera shake correction, and a clock circuit 87 for recording photographing time are connected via a system bus 88. Consists of.

  The system controller 80 controls the entire control unit 53 and uses the RAM 82 as a work area. In the ROM 81, a program for controlling the camera unit 51 and a program for executing recording control and reproduction control of video signals and audio signals are written.

  An operation input unit 90 connected to the operation input interface 83 includes a mode switching key for switching between a shooting mode and another mode such as a playback mode, a zoom adjustment key, an exposure adjustment key, a shutter key, and a moving image shooting key. A plurality of keys such as a display adjustment key in the display unit 91 are provided. The operation input interface 83 transmits an operation signal from the operation input unit 90 to the system controller 80. The system controller 80 determines which key is operated on the operation input unit 90 and performs control processing according to the determination result.

  The display unit 91 connected to the display control unit 84 is configured by, for example, an LCD (Liquid Crystal Display) or the like, and is read from the video signal from the camera unit 51 or the disk drive 200 under the control of the system controller 80. Display the video signal.

  The memory card interface 85 writes the compressed data from the encoding / decoding unit 70 to the memory card 92. The memory card interface 85 reads the compressed data from the memory card 92 and supplies the compressed data to the encoding / decoding unit 70. The clock circuit 87 generates time information representing year, month, day, hour, minute, second and the like.

  The angular velocity detector 86 is a gyroscope that detects an angular velocity applied to the imaging device 50 from the outside. The angular velocity information [ω = (θ / second)] from the angular velocity detector 86 is reported to the system controller 80 at predetermined intervals. When the integral value [θ] from the start of recording exceeds a predetermined value (for example, 5 °), a flag is set in STB_LIM of the additional recording information as exceeding the limit of camera shake correction. Note that ω is + ω when shifted to the right from the center of the screen, and −ω when shifted to the left, and has limit values in both positive and negative directions.

  FIG. 9 shows an exemplary configuration of the disk drive 200. The disk drive 200 is connected to the spindle motor 201 that rotates the recordable optical disk 214, the spindle motor driver 207 that generates a drive signal for driving the spindle motor 201 and supplies the drive signal to the spindle motor 201, and the recordable optical disk 214. An optical pickup 213 that irradiates a laser beam and receives a laser beam reflected by the recordable optical disc 214, a sled motor 202 that moves the optical pickup 213 in the radial direction of the recordable optical disc 214, and an illustration inside the optical pickup 213 And a two-axis control unit 205 that controls driving of the two-axis mechanism that is not performed.

  The disk drive 200 also includes an RF unit 203 that processes the output signal of the optical pickup 213, a signal processing unit 204 that generates various control signals based on the output of the RF unit 203, and a program stored in advance in a ROM (not shown). The microcomputer 208 that controls the entire disk drive 200 according to the above, the non-volatile memory (flash memory) 209 that can be rewritten and retains the written data even after the power is turned off, and the outside, that is, the recording / reproducing described above. And an interface (I / F) unit 212 that controls communication with the processing unit 52.

  In the spindle motor 201, a disk table for mounting the recordable optical disk 214 is integrally attached to the drive shaft. Based on the drive signal supplied from the spindle motor driver 207, the drive shaft is fixed at, for example, a linear velocity ( The recordable optical disk 214 mounted on the disk table is rotated by being driven to rotate at CLV (Constant Linear Velocity) or constant angular velocity (CAV: Constant Angular Velocity). The spindle motor driver 207 outputs a pulse corresponding to the rotation of the drive shaft. This pulse is supplied to the audio signal processing unit 65 as a reference signal X synchronized with the period of periodic noise.

  Although not shown, the optical pickup 213 has a laser diode as a laser light source, a laser diode driver for driving the laser diode, and a laser beam emitted from the laser diode as a laser beam. It has an optical system for condensing, and a photodetector for detecting the laser light reflected from the recordable optical disk 214. The optical system is driven biaxially in the disk surface direction and the disk radial direction based on the drive signal supplied from the biaxial control unit 205, and the focus and tracking are controlled.

  For example, the optical system irradiates the recordable optical disk 214 by dividing one laser beam into one zero-order light and two first-order lights using a grating or the like. The photodetector is composed of, for example, two two-divided detectors each having a light-receiving surface divided into two, and a four-divided detector having four light-receiving surfaces, and each of the two primary lights is received by the two-divided detector. At the same time, one reflected light of the 0th order light is received by the four-divided detector, and detection signals corresponding to the light receiving surfaces are respectively output.

  Detection signals from the respective divided light receiving surfaces of the photodetector output from the optical pickup 213 are supplied to the RF unit 203. The RF unit 203 performs amplification processing on each supplied detection signal, performs predetermined calculation between the signals, generates a focus error signal and a tracking error signal, and outputs from the recordable optical disc 214. A reproduction RF signal corresponding to the reflected laser beam is generated. The focus error signal and the tracking error signal are supplied to the signal processing unit 204. The reproduction RF signal is demodulated in a predetermined manner to be a reproduction signal, which is supplied to the signal processing unit 204.

  Further, the RF unit 203 generates a light amount control signal for controlling the amount of laser light emitted from the optical pickup 213 to the recordable optical disc 214 based on the signal supplied from the optical pickup 213, and the optical pickup 213. To the laser diode driver. The RF unit 203 generates a light amount control signal for controlling the amount of laser light emitted from the optical pickup 213 to the recordable optical disk 214 to be constant during reproduction, and outputs it from the signal processing unit 204 during recording. The level of the light quantity control signal is controlled according to the recording signal to be recorded.

  During reproduction, the signal processing unit 204 A / D-converts the supplied reproduction signal based on the control of the microcomputer 208 and temporarily stores it in the storage unit 206. Then, the storage unit 206 is used to decode the recording code of the reproduction data and the decoding process of the error correction code. The decoded reproduction data is read from the storage unit 206 at a predetermined timing and supplied to the recording / reproduction processing unit 52 via the I / F 212.

  Further, at the time of recording, the signal processing unit 204 temporarily stores the recording data supplied via the I / F 212 in the storage unit 206, and uses the storage unit 206 to perform error correction encoding processing and recording code on the recording data. The process is applied. Then, a recording pulse is generated based on the recording data subjected to these processes, and is supplied to the optical pickup 213.

  Further, the signal processing unit 204 generates management information of data to be recorded on the recordable optical disc 214 and temporarily stores it in the flash memory 209, for example. Then, for example, according to an instruction from the microcomputer 208, the management information stored in the flash memory 209 is referred to, and each unit of the disk drive 200 is controlled to execute various reproduction methods such as random reproduction and shuffle reproduction.

  The focus error signal and tracking error signal supplied to the signal processing unit 204 are A / D converted, subjected to predetermined signal processing, and supplied to the two-axis control unit 205. The biaxial control unit 205 controls the operation of the biaxial mechanism of the optical pickup 213 based on the supplied signal. Further, the biaxial control unit 205 generates a drive signal for moving the optical pickup 213 to the designated track position on the recordable optical disc 214 in accordance with a command from the microcomputer 208 and supplies the drive signal to the sled motor 202.

  Next, a fifth embodiment of the present invention will be described with reference to FIG. In the fifth embodiment of the present invention, for example, when video and audio are reproduced from a recording medium, periodic noise included in the audio signal during recording included in the reproduced audio signal is reduced. It is. In the following description, parts common to those in FIG. 7 described above are denoted by the same reference numerals and description thereof is omitted.

  Consider a case where a recording medium recorded in a state where periodic noise mixed in the main audio signal during recording is not reduced is reproduced by a device having a disk drive similar to the disk drive shown in FIG. The periodic noise mixed in the main audio signal is periodic noise generated by the rotation of the spindle motor that rotates the optical disk. For example, in the case of DVD video, the rotation cycle of the spindle motor is determined in advance based on the standard. Therefore, even when the reproduction is performed by a device different from the recording time, the rotation frequency of the spindle motor is the same as the device at the time of recording. Conceivable.

  That is, even if the device used for recording is different from the device used for playback, the period of the mixed periodic noise is considered to be the same. Even when the device used at the time of reproduction is different, periodic noise can be reduced based on the reference signal X supplied from the spindle motor driver of the device used at the time of reproduction.

  Here, for example, a reproduction operation of an audio signal recorded on a disc-shaped recording medium such as an optical disc will be schematically described. During reproduction, the optical disk 214 is mounted on a disk table attached to the drive shaft of the disk drive 200, and the optical disk 214 is rotated by the spindle motor 201. At this time, a reference signal X is output from a spindle motor driver 207 that controls the rotation of the spindle motor 201.

  A detection signal output based on the reflected light of the light irradiated to the optical disk from the optical pickup 213 of the disk drive 200 is subjected to predetermined processing by the RF unit 203 and supplied to the signal processing unit 204 as a reproduction signal. The reproduction signal is subjected to predetermined processing by the signal processing unit 204 and supplied to the recording / reproduction processing unit 52 via the I / F 212.

  The compression-encoded and multiplexed reproduction data supplied to the recording / reproduction processing unit 52 is supplied to the encoding / decoding unit 70, and the encoding / decoding unit 70 uses the buffer memory 71 to reproduce the reproduction data. The video signal, the audio signal and the additional recording information are separated and decoded. The decoded audio signal is supplied to the audio signal processing unit 65, and the reference signal X supplied from the disk drive 200 by the noise reduction circuit 1, 1 ′, or 1 ″ is used in the first embodiment described above. The periodic noise reduction processing described in the second or third embodiment is performed, and the periodic noise mixed in the reproduction data is reduced and output.The audio signal in which the periodic noise is reduced is a system signal. The sound is supplied to the sound output unit 89 connected via the bus 88 and output from the speaker 93 as sound.

  In the above description, the method for reducing periodic noise by a spindle motor for rotating a disk-shaped recording medium such as an optical disk has been described. However, this is not limited to this example. For example, when a magnetic tape is used as a recording medium, the present invention can also be applied to the case where periodic noise generated by the rotation of a rotating drum is reduced. For example, the reference signal X can be generated based on a PG (Pulse Generator) signal or an FG (Frequency Generator) signal output in synchronization with the rotation of the rotating drum. In other words, the periodic noise reduction circuit according to the present invention is applied to the case where a rotation mechanism that rotates the recording medium at a constant period is provided and the period information of the signal that controls the rotation of the rotation mechanism can be acquired. Is possible.

  The periodic noise reduction circuit according to the first, second and third embodiments of the present invention can also be applied to a device which records only sound. Specifically, the present invention can be applied to an apparatus in which the optical block 60, the camera control unit 61, the signal converter 62, and the imaging signal processing unit 63 are excluded from the circuit shown in FIG. Of course, the present invention can also be applied to an apparatus that only reproduces audio.

It is a block diagram which shows the structure of an example of the periodic noise reduction circuit by the 1st Embodiment of this invention. It is a basic diagram which shows an example of the signal waveform of each part in the periodic noise reduction circuit by the 1st Embodiment of this invention. It is a block diagram which shows the structure of an example of the periodic noise reduction circuit by the 2nd Embodiment of this invention. It is a basic diagram which shows an example of the signal waveform of each part in the periodic noise reduction circuit by the 2nd Embodiment of this invention. It is a block diagram which shows the structure of an example of the periodic noise reduction circuit by the 3rd Embodiment of this invention. It is a basic diagram which shows an example of the signal waveform of each part in the periodic noise reduction circuit by the 3rd Embodiment of this invention. It is a block diagram which shows the structure of an example of the imaging device by the 4th Embodiment of this invention. It is a block diagram which shows the structure of an example of an audio | voice signal processing part. It is a block diagram which shows the structure of an example of a disk drive. It is a block diagram which shows the structure of an example of the imaging device by the 5th Embodiment of this invention. It is a block diagram which shows the structure of an example of the periodic noise reduction circuit using an adaptive filter. It is a block diagram which shows the structure of an example of an adaptive filter.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Input terminal 11 Buffer memory 12 Correlation calculating part 13 Correlation determination part 14 Adder 15 Multiplier 16 Adder 17 Noise memory 18 Adder 19 Input terminal 20 Address generation part 21 Output terminal 30 Limiter 31, 32 Selector 33 Initial noise taking determination Part 50 imaging device 65 audio signal processing part

Claims (12)

A storage unit for storing at least one period of the periodic noise period of the input audio signal;
Correlation between one period of the periodic noise period of the input audio signal and the audio signal of one period one period before the input audio signal stored in the storage unit. A correlation calculation unit to calculate,
A correlation determination unit that determines a gain coefficient for each cycle of the cycle based on the correlation calculated by the correlation calculation unit;
Signal processing characterized in that the audio signal read from the storage unit is multiplied by the gain coefficient for each cycle and accumulated, and the accumulated signal is subtracted from the inputted audio signal. circuit. The signal processing circuit according to claim 1,
The signal processing circuit, wherein the correlation determination unit determines the gain coefficient based on a threshold value for a correlation coefficient indicating the correlation. The signal processing circuit according to claim 1,
A limiter for limiting the level of the audio signal to a predetermined level;
The signal processing circuit according to claim 1, wherein when the level of the input audio signal is limited by the limiter, the correlation determination unit determines the gain coefficient corresponding to the period as 0. The signal processing circuit according to claim 1,
While holding the audio signal stored in the storage unit at a predetermined timing, monitor the correlation calculated by the correlation calculation unit,
When it is determined that the correlation between the input audio signal and the audio signal held in the storage unit is continuously high for a predetermined period, the audio signal held in the storage unit is accumulated. A signal processing circuit characterized in that the initial value of the audio signal is set. The signal processing circuit according to claim 4,
The signal processing circuit according to claim 1, wherein the timing is a start-up time. Store at least one period of the periodic noise period of the input audio signal in the storage unit,
Correlation between one period of the periodic noise period of the input audio signal and the audio signal of one period one period before the input audio signal stored in the storage unit. Calculate
Based on the calculated correlation, a gain coefficient is determined for each of the periods,
Signal processing characterized in that the audio signal read from the storage unit is multiplied by the gain coefficient for each cycle and accumulated, and the accumulated signal is subtracted from the inputted audio signal. Method. An imaging unit that images light from a subject and outputs a video signal;
A sound collection unit built in the housing that collects sound and outputs a sound signal;
An audio signal processing unit that performs signal processing on the audio signal output from the audio pickup unit;
A recording unit having a rotating mechanism that rotates in a predetermined cycle, and records the video signal output from the imaging unit and the audio signal output from the audio signal processing unit on a recording medium;
The audio signal processing unit is
A storage unit for storing at least one period of a periodic signal synchronized with the rotation period of the rotation mechanism of the input audio signal;
Calculate the correlation between the period of the periodic signal of the input audio signal and the audio signal of one cycle one period before the input audio signal stored in the storage unit. A correlation calculation unit,
A correlation determination unit that determines a gain coefficient for each cycle of the cycle based on the correlation calculated by the correlation calculation unit;
The audio signal read from the storage unit is accumulated by multiplying the gain coefficient for each period, and the accumulated signal is subtracted from the input audio signal and output. An imaging device. A video signal output step of imaging light from a subject in an imaging unit and outputting a video signal;
A sound signal output step of collecting sound and outputting a sound signal with a sound collecting unit built in the housing;
An audio signal processing step for performing signal processing on the audio signal output from the audio signal output step;
Recording the video signal output from the video signal output step and the audio signal output from the audio signal processing step on a recording medium with a recording unit having a rotation mechanism that rotates at a predetermined period;
The audio signal processing step includes
Storing at least one period of the periodic signal synchronized with the rotation period of the rotation mechanism of the input voice signal in the storage unit;
Calculate the correlation between the period of the periodic signal of the input audio signal and the audio signal of one cycle one period before the input audio signal stored in the storage unit. And
Based on the calculated correlation, a gain coefficient is determined for each of the periods,
The audio signal read from the storage unit is accumulated by multiplying the gain coefficient for each period, and the accumulated signal is subtracted from the input audio signal and output. An audio signal processing method for an imaging apparatus. A sound collection unit built in the housing that collects sound and outputs a sound signal;
An audio signal processing unit that performs signal processing on the audio signal output from the audio pickup unit;
A recording unit having a rotation mechanism that records the audio signal output from the audio signal processing unit and rotates at a predetermined period;
The audio signal processing unit is
A storage unit for storing at least one period of a periodic signal synchronized with the rotation period of the rotation mechanism of the input audio signal;
Calculate the correlation between the period of the periodic signal of the input audio signal and the audio signal of one cycle one period before the input audio signal stored in the storage unit. A correlation calculation unit,
A correlation determination unit that determines a gain coefficient for each cycle of the cycle based on the correlation calculated by the correlation calculation unit;
A recording apparatus characterized in that the audio signal read from the storage unit is multiplied by the gain coefficient for each cycle and accumulated, and the accumulated signal is subtracted from the inputted audio signal. . A sound signal output step of collecting sound and outputting a sound signal with a sound collecting unit built in the housing;
An audio signal processing step for performing signal processing on the audio signal output from the audio signal output step;
Recording the audio signal output from the audio signal processing step with a recording unit having a rotation mechanism that rotates at a predetermined cycle;
The audio signal processing step includes
Storing at least one period of the periodic signal synchronized with the rotation period of the rotation mechanism of the input voice signal in the storage unit;
Calculate the correlation between the period of the periodic signal of the input audio signal and the audio signal of one cycle one period before the input audio signal stored in the storage unit. And
Based on the calculated correlation, a gain coefficient is determined for each of the periods,
A recording method characterized by accumulating the sound signal read from the storage unit by multiplying the gain coefficient for each period and subtracting the accumulated signal from the input sound signal. . A reproduction unit having a rotation mechanism for reproducing an audio signal and rotating at a predetermined period;
An audio signal processing unit that performs signal processing on the audio signal output from the reproduction unit;
An audio output unit that outputs the audio signal output from the signal processing unit,
The audio signal processing unit is
A storage unit for storing at least one period of a periodic signal synchronized with the rotation period of the rotation mechanism of the input audio signal;
Calculate the correlation between the period of the periodic signal of the input audio signal and the audio signal of one cycle one period before the input audio signal stored in the storage unit. A correlation calculation unit,
A correlation determination unit that determines a gain coefficient for each cycle of the cycle based on the correlation calculated by the correlation calculation unit;
The audio signal read from the storage unit is accumulated by multiplying the gain coefficient for each period, and the accumulated signal is subtracted from the input audio signal and output. Playback device. An audio signal reproduction step of reproducing an audio signal by a reproduction unit having a rotation mechanism that rotates at a predetermined period;
An audio signal processing step for performing signal processing on the audio signal output from the audio signal reproduction step;
Outputting the audio signal output from the audio signal processing step,
The audio signal processing step includes
Storing at least one period of the periodic signal synchronized with the rotation period of the rotation mechanism of the input voice signal in the storage unit;
Calculate the correlation between the period of the periodic signal of the input audio signal and the audio signal of one cycle one period before the input audio signal stored in the storage unit. And
Based on the calculated correlation, a gain coefficient is determined for each of the periods,
The audio signal read from the storage unit is accumulated by multiplying the gain coefficient for each period, and the accumulated signal is subtracted from the input audio signal and output. How to play.

Images